Playwright MCP Quickstart
Enables AI agents to control a local browser via Playwright's accessibility tree, allowing automated web interactions and testing through natural language commands.
README
Playwright MCP Quickstart Template 🤖🎭
Welcome to the Quickstart Template for running Playwright with the Model Context Protocol (MCP). This repo contains everything you need to let AI agents (like Claude Desktop or Cursor) control a local browser using Playwright's accessibility tree.
As seen on JKC Tutorials.
🚀 Prerequisites
- Node.js 20 or newer installed on your machine.
- An MCP client installed (e.g., Claude Desktop or Cursor).
⚙️ Installation
- Clone this repository.
- Install the dependencies:
npm install - Install the required Playwright browsers:
npm run setup
🔌 Connecting to Claude Desktop
To give Claude Desktop control over Playwright, you must update Claude's configuration file.
- Open the Claude Desktop configuration file:
- Mac:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
- Mac:
- Copy the contents of the
claude_desktop_config.jsonfile provided in this repository and paste it into Claude's config file. - Restart Claude Desktop.
Note: You do not need to run npm run start-mcp manually if you configure Claude Desktop this way; Claude will spin up the server automatically when needed.
🧪 Running Your First AI Test
Once Claude Desktop is restarted, open a new chat in Claude and copy/paste this exact prompt:
"Use Playwright MCP to navigate to https://www.saucedemo.com/. Log in using the username 'standard_user' and password 'secret_sauce'. Add the backpack to the cart, navigate to the cart, and verify that the backpack is present."
Watch as Claude opens a headed browser, reads the accessibility tree, and executes the entire flow without a single line of traditional test code!
Want more tutorials on AI and Test Automation? Subscribe to JKC Tutorials on YouTube.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。